This proposal is based on the premise that detection of risk factors for breast cancer is hampered by the fact that breast cancer is really a collection of diseases with distinct etiologies. If we are able to define sub-types of breast cancer that we are convinced are etiologically heterogeneous then we can conduct studies to discover novel genetic risk factors in a more effective, statistically powerful way by searching for new risk factors within the sub-types. Our thesis is that unique insights for establishing etiologically distinct sub-types can be obtained by examining the correlations of molecular characteristics of tumor pairs in women with contralateral breast cancer. We will use archival tumor and normal tissue from a completed case-control study of breast cancer (the Cancer and Steroid Hormone (CASH) Study) that involved follow-up of the cases for the occurrence of a contralateral cancer to serve as a prototype investigation of these ideas. The study will follow a logical sequence of steps. In Aim 1 we will profile pairs of contralateral tumors on the basis of expression patterns i a custom panel of genes in pathways implicated in breast cancer and global micro RNA expression in order to examine the concordance of these profiles. This information will be used to define the sub-types with the greatest evidence for etiologic heterogeneity. In Aim 2 tumor specimens from a larger collection of incident cases of breast cancer will be similarly profiled and classified into the sub-types defined in the first aim. These sub-types will be compared with respect to existing, known risk factors for breast cancer that are available in the CASH Study to determine if the new classification system demonstrates that the impact of selected risk factors are focused primarily on individual tumor sub-types. Finally, in Aim 3 normal tissue from the cases will be used to obtain information on a custom panel of germ-line variants, and these will be explored to identify variants that are unique to or concentrated in one or other of the tumor sub-types. The study will make use of existing data and specimens from an archival case-control study to conduct a prototype investigation of a research strategy that has the potential to improve greatly the yield from future epidemiologic studies in both breast cancer and other cancer sites.